Tech behind new V10 M5's ignition system.

[Shaun]

Been lots of posts recently linking to sites that detail the technology in the new M5. IMO the ignition system is the most revolutionary thing about it since it allows accurate and direct measurement of 2 important areas (1 essential) and AFAIK it has never been run on any production car or race car. Below is an article from Racecar Engineering October 2003. There are many areas I need some clearing up on and my comments and questions will be in [] brackets. If someone with knowledge of electronics or an electrical engineer can help, TIA. If any diagram needs better resolution or clarity let me know and I'll take them separately and full frame.

===========

Keith Howard

During the 1980s researchers from the Institut fur Physikalische Elektronik at the Universtiy of Stuttgart published a series of papers about a new form of IC engine ignition system they termed 'breakdown' ignition (aka plasma ignition). Whereas conventional inductive ignition produces an ignition event lasting a couple of milliseconds and a capacitive discharge one lasting perhaps 100 microseconds, the oscillatory spark current in the breakdown ignition system reduced to near zero within 200 nanoseconds. - 10,000 and 500 times faster respectively. Although the breakdown system showed promise in respect of improved lean mixture ignitability, insensitivity to plug fouling and lowered fuel consumption and emissions, it was never employed in production.

Scene change to the late 1980s and a change meeting between two students at MIT: Ed VanDyne, who at the time was in charge of the MIT racing team running Formula Vees, and Stefan Pischinger, who was working on a PhD thesis investigating the lean limit performance of a range of different spark plug designs for Bosch. (Pischinger is now CEO of engineering consultancy FEV in Aachen, Germany, the company founded by his father.) Pischinger was running a single cylinder test engine with a square section cylinder that allowed combustion events to be filmed through its transparent side walls, a mirror arrangement allowing the flame kernel to be viewed from two directions at right angles so that flame growth could be visualized in three dimensions.

"Every once in a while over a six month period he'd come in at the same time we were there and we'd watch one of his movies," recalls VanDyne. "There was one night when he had a dramatically different spark plug, from the look of it in the movie, and it had dramatically different combustion results. The frame where the arc occurred was very bright compared to all the other movies I'd seen, and the flame growth was remarkable. One frame there was nothing, next frame the flame kernel was already growing fast enough that two or three frames later it was off the screen. In the other movies there would be 10 to 12 frames before the flame even grew to the point where it engulfed the spark plug. This plug did that by the second frame after the spark!"

"The growth was so dramatic that I decided to find out more about what Stefan was doing. It turned out that he had built capacitance into the spark plug to act as an energy store. When the plug fired, the capacitor dumped a very high surge current across the gap that made a very hot spark." (Nology's aftermarket HotWires ignition leads exploit this same idea by placing an external capacitor as close as possible to the spark plug, which discharged through the gap once the breakdown voltage has been achieved.)"

The design of the spark plug itself was a vital factor too. Whereas a conventional J-gap plug has an earth electrode which curves over the centre electrode, blocking flame kernel growth to some degree (sometimes considered a benefit as it shields the developing kernel from mixture turbulence), the spark plug that had given such an impressive account of itself on the combustion movie was a projected surface gap type (Figure 1)



which relies on electrical conduction across the surface of the ceramic insulator to one or more short ground electrodes arrayed around it. [Note: there are actually four or more short ground electrodes on the plug, not just one like shown in figure 1]

To VanDyne's surprise, Stefan Pischinger didn't consider that the faster kernel growth provided y the combination of breakdown ignition and a surface gap plug was of practical significance. "I looked into it," says VanDyne "and realized that sure, the breakdown ignition system used ahead of the spark plug wasn't practical. What you needed to be able to do was turn that high energy pulse on and off at will. So I went over to the MIT strobe light lab and talked to the head technician. He and I went over how a strobe light circuit worked and I asked him if I could combine an ignition circuit with a strobe circuit. He said he didn't see why not. So in 1990, two 1940s patents - capacitive discharge ignition and a strobe light were combined to create something patentable in its own right (Figure 2). I went to the MIT technology licensing office who offered to file the patents and license them back to me." And so Adrenaline Research Inc was born as a spin-off company from MIT, initially by Ed VanDyne and two colleagues.

The short high-current pulse used in the Pischinger ignition lasted 100-500 nanoseconds. Out high-current pulse from the strobe circuit, has a long duration of about three microseconds, with a peak current of 100-300 amps. So the pulse train is slightly different but both create a bright plasma discharge. In our system we have about 100 microsecond duration overall because we still have a capacitive discharge ignition firing the coil. As soon as the gap has broken down it provides the conductive path for the high power from the strobe circuit to be dumped in three microseconds, then the low current, about 500 milliamps continues to flow from the coil for another 100 microseconds. The plasma discharge on the project surface gap plug promotes faster flame kernel development because the high current on the insulator surface dissipates 4 times as much electrical heat energy into the combustible mixture. The best analogy is that of an arc welder – when you turn up the current, you turn up the heat.

Although this short duration high energy spark has direct benefits in respect of improved lean limit performance and increased engine power, this isn’t what justifies the name Smartfire for the Adrenaline ignition system. The indirect benefit of this brief ignition even is that it facilitates effective monitoring of the ionization within the cylinder, which allows misfire and knock detection to be performed more reliably than with the conventional methods and cylinder by cylinder. Real time closed loop control ignition timing and fuelling is also made possible, again on a bespoke basis of each cylinder. These capabilities are what make VanDyne’s ignition technology ‘smart’.

Ionisation occurs within the combustion process when free radicals, intermediate chemical species are formed which carry an electrical charge. These provide a comductive path across the spark gap during the combustion even, the conductivity of which is related to the progress of combustion. In the Smartfire circuit this is monitored by placing a relatively low voltage (100-300 volts) across the secondary capacitor (which is also used, at higher voltage, to store charge for the high energy plasma ignition) and measuring the gap current via the resistor.

Exactly how closely the combustion and ionization are related you can judge from Figure 3, which superimposes graphs of cylinder pressure and the equivalent ionization current (monitored as a voltage across R). There are two peaks in the ionisation curve, the first of which occurs shortly after the spark. This is related to the combustion process when the flame touches the tip of the plug. A second smaller peak in the ionisation curve occurs later and is closely related to the temperature changes within the cylinder. As a result, the second peak of the ionisation curve occurs at the same crank angle position, on average, as peak cylinder pressure. This indirect determination of where peak cylinder pressure occurs is what allows Smartfire to execute closed loop, cylinder by cylinder control of fuelling and ignition timing. Detection of misfire is almost trivially simple: the ionisation signal simply ‘flatlines’.

In fact the ionisation tracks temperature within the cylinder so closely that if the cylinder is knocking (figure 4), telltale fluctuations are clearly visible in the ionisation signal at around 6 kHz (or sometimes the second harmonic). This oscillation can be detected by digital signal processing within Smartfire, which flags when knock is detected. Again, this is achieved on a cylinder by cylinder basis.



It’s worth a little break for a bit of a historical note: Saab used a combination of capacitive ignition and ionisation monitoring on its road cars some years ago but has begun to phase this out and is in the process of reverting to inductive ignition. The overriding need in modern road cars is to prevent misfire, both to protect the catalytic converter from damage and enable the vehicle to pass increasingly stringent emissions tests. In this respect, inductive ignition is better than conventional capacitive discharge – a shortfall that Smartfire makes up though its addition of the strobe circuit.

Catalyst protection is irrelevant in most racing formulae, of course, but the Smartfire system’s ability to detect misfire and knock and to provide closed loop adjustment of ignition and fuelling for each cylinder, make it just as relevant on the race circuit as the street. “I never thought misfire would remotely be an issue in racing engines, but it turns out ir really is,” says VanDyne. “I’m hearing more and more from racecar builders that they’re having full load misfiring because of the they’re now fuelling the engines.

“I can’t describe some of the proprietary fuelling methods but it’s public knowledge that Audi has used its FSI (Fuel Stratified Injection) technology at Le Mans. Others are trying to follow suit but it’s a very difficult thing to achieve. FSI was originally designed for low emissions and high fuel economy in street cars, but its fuel economy advantage can be significant in racing too. Audi won the first time at Le Mans without the technology. When they came back the second year, the reason they went a whole lap further between pit stops was the fuel economy advantage of the FSI. So Audi’s dominance in 2001 and 2002 was partly due to the fuelling technology. But these advanced fuelling methods have the disadvantage that they can cause misfires under conditions where in the past, racecars didn’t misfire. Where fuelling issue are becoming critical, normal ignition systems aren’t keeping up. That’s where we have an advantage. “

“At these high engine speeds we don’t even attempt misfire prevention. It’s not an issue that the misfire comes out of the exhaust, whereas in a road car it is – there can be conditions where we may do misfire prevention (a second spark) 50 degrees after the first ignition just to attempt to prevent that slug of fuel and air getting to the catalyst. For racecar applications misfire detection is useful, at least as a diagnostic tool. We have performed misfire detection up to 18,000rpm, so we really can do it fast enough for today’s high speed racing engines.”

“We wouldn’t necessarily use high energy ignition in a racecar application, although some will like the extra power of the faster flame kernel development. We’ve done experiments at high engine speeds which show that, even though there’s a rich mixture, you can retard the timing by 4 to 6 degrees and get, typically, a 0.5 to 1 percent increase in power. The later you start the burn before TDC the less work you lose pushing against the piston on the upstroke. That’s an advantage before you do any ionisation monitoring.” [Note: The only way you get more power by retarding ignition at minimum advance for best torque, is if the burn time is compressed so that peak cylinder pressure occurs at the same (optimum) degree after TDC. Least average pressure before TDC and highest average pressure after TDC up to a certain point yields best power. Retarding alone will lose power.]

“The potential problem is, at full load the high energy discharge will erode the spark plugs about five times faster than normal. I don’t like some spark plug manufacturer’s because they use inexpensive materials, not even Inconel, in the racing spark plugs. We need Inconel to get the durability our high energy sparks require, otherwise even over a race distance the plug can be over-eroded. Fortunately some of the manufacturers make their 10mm F1 plugs with Inconel electrodes so we can avoid problems with plug life. Also we have an incredibly good machinist close by who can rework spark plugs for us. We’ve shown horsepower improvement by cutting the side electrodes off a Bosch plug and inserting Inconel pins to turn it into a surface gap plug.”

To date Smartfire has only been used in the dyno cells of racing engine manufacturers, but an on-car system is no on the horizon as a result of the licensing deal Adrenaline has signed with Motorola to develop the technology for road car use.

“We had hoped to get a racing partner to fund making an in-car hardened system. Teams liked the concept but even in F1 they are very conservative. They want it handed to them in the final form that goes in the racecar. They certainly don’t want to pay us to develop it. But together with Motorola we are developing the system for the automotive world and the plan is to use this module as the first generation in-car racing system. It will be subject to much more rigorous testing than we’ve been able to put our existing box through, and it will be more compact. Our current controller measures about 3 x 5 x 5 inches and is not ideal for a high vibration environment, so it’s not really suitable for fitting to a racecar. We have to get it out quickly and with all the features needed for dyno test cells, so we sell this as the instrumentation product and expect to make the in-car version together with Motorola."

“OEM production is slated to start in time for the OBD2 100 percent misfire standard that comes effective in California in 2005. So we have about 15-18 months development time remaining for that. For the racecar people the unit could be available earlier because we will have pre-production prototypes this fall. They become, as far as I’m concerned, in-car racing test systems. “

Smartfire’s ionisation monitoring will be put to work from the outset, initially for knock detection. “The first customer we expect to test with primarily needs the knock detection. They will remove their block transducer and use our system as the input to their knock algorithms. Currently they get significant ignition retard -8-10degrees because the transducers pick up valvetrain and other noises. They know the engine can knock, and they’ve calibrated it fairly aggressively, they also know that if it does knock it’s only on one or two cylinders. Because all the cylinders are retarded when knock is detected, their current method is really not working. Whereas we can do knock detection on a cylinder by cylinder basis, we won’t flag all the cylinders unless all the cylinders are knocking. They think they’ll gain a significant horsepower advantage from being able to retard the cylinders individually.”

The next stage – real time closed loop control of ignition timing and fuelling based on the ionisation data – is already feasible. “We have a CAN bus on which we can output critical information that allows closed loop tuning. The algorithm is already written and we’ve establishes that the system works with many different types of combustion chamber. Determining the ideal ignition timing for each cylinder is a no brainer because the ionisation signal identifies when peak cylinder pressure is developed. It is the ideal air fuel ratio that takes a bit more of an algorithm (figure 5).




”The power advantage is very variable between different engines – it could be as little as less than 0.5 percent, but some F1 engine manufacturers can have a 3 percent output variation engine to engine. What they are not able to do is diagnose why this occurs when they build them to such exacting tolerances. We are looking to show them where the deficiency lies by being able to do a tune up on each cylinder. We think our system has the capability to null out that 3 percent variation, or at least reduce it to one percent. So it may not raise the power of the best engines at all but it could take the worst engines up two percent, so they don’t have to scrap them. Or they could use our system to diagnose the problem and fix it.”

“The big value, as in a street car, is that we can measure combustion effects without intruding on the cylinder in any new way – our technology is completely non-invasive. You can plug Smartfire on to any race engine and monitor its combustion characteristics whereas an instrumented engine, with in-cylinder pressure sensors, could never be fitted in a racecar. That’s the great benefit of this system.”

===========


Was supposed to type out questions but found it too interruptive to typing flow. Will post soon.

From the pictures I posted about in the other M5 thread, you can see that it still has two sets of pre and post cat oxygen sensors, as well as twin EGT probes. So it still has redundant systems for monitoring AFR and combustion temps – though the oxygen probably primarily for monitoring cat efficiency. VanDyne and Pischinger must be extremely rich men now – esp since F1 and BMW are taking up their patented system.